fix off-by-one subtree check in prefixMap iterator remove

[rootvector2]

PrefixRangeEntrySet.EntryIterator.remove() marks the iteration finished whenever lengthInBits >= subtree.bitIndex, but a subtree root whose bitIndex equals the prefix length in bits is a branch node rather than the only matching entry (PrefixRangeEntrySet.iterator() uses the strict > for that same test and only then returns a SingletonIterator), so removing andes from prefixMap("and") over the keys and, andes, android, androids ends the iteration after android and silently leaves androids in the trie, and subtree() returns null once no entry with the prefix is left so the following subtree.bitIndex read throws; found with a randomized differential test of prefix-view iterator removal against a filtered TreeMap, where 200000 random cases give about 3100 NullPointerExceptions, 340 truncated iterations and 190 tries whose final content differs from the reference before this change, and 600000 cases are clean after it.

• Read the contribution guidelines for this project.
• Read the ASF Generative Tooling Guidance if you use Artificial Intelligence (AI).
• I used AI to create any part of, or all of, this pull request. Which AI tool was used to create this pull request, and to what extent did it contribute? Claude Code: it ran the differential test that located the defect and drafted the patch and the regression test, both of which were reviewed and verified locally.
• Run a successful build using the default Maven goal with mvn; that's mvn on the command line by itself.
• Write unit tests that match behavioral changes, where the tests fail if the changes to the runtime are not applied. This may not always be possible, but it is a best practice.
• Write a pull request description that is detailed enough to understand what the pull request does, how, and why.
• Each commit in the pull request should have a meaningful subject line and body.